Re: How do Banking Systems Work

On 13 Dec 1995 18:49:14 GMT, frank surewood <franks_at_qnet.com> wrote:

>I am running into problems performing selects with my very large Oracle 7
>database (a million records). How does the banking system or similar
>high volume applications handle this problem? Example, how does my ATM
>respond so quickly when the select must be in the same scale as my
>application to find my records? Is the answer big and better hardware?
>
>Frank Surewood
>

Well, more powerful hardware can't hurt - esp a move to parallelism. However, it seems to me that you could have an optimisation problem as 1M row tables ain't so very big for a banking system. My current project uses 6 processors on the development system (14 for the production machines) and multiple disks (striped). We also partition our tables and maintain metadata to keep track of it all. I don't really know much about the hardware side of things other than what I pick up from the configuration chappies on site but I understand that when the whole caboodle is put together it will run like doos off a shiney shouvle. At the moment a simple _select from group by_ runs inside two minutes if there is not much else going on with the machine.

I have no indication of your experience so I hope I'm not being patronising when I suggest explain plan.

Hope this helps.

Steve

Steve Roach                              Tel: (44) 0973 162755
Stercon Systems Ltd.                 steve_at_stercon.demon.co.uk